Anthropic Accuses Zhipu of Distilling Claude for GLM-5.2
AccBalanced · x · 2026-07-16
Speaking at the Aspen Security Forum, Anthropic's head of national security policy Tarun Chhabra accused 智谱 of distilling Claude and OpenAI models to train GLM-5.2. He also previously named DeepSeek, Alibaba, Moonshot, and MiniMax.
The post highlighted several key takeaways: Anthropic stated it continuously detects and bans distillation accounts on a scale of millions per week. Chhabra noted that GLM is "currently one of the most advanced Chinese models on the market," posing significant cybersecurity challenges. Additionally, Anthropic might expand access to Mythos to help cyber defenders "fight fairly."
Related event: Anthropic Accuses Zhipu of Distilling Claude Models(3 posts)→
More from Models
- MissionReport benchmark: only 3 of 540 AI-generated government reports pass all criteria — Exp_Mark · 2026-10-27
- Local beats cloud: Qwen 3.8 FN handles private student data as well as frontier models — dr_lm · 2026-10-27
- Aidan Gomez: model's CoT said no tool calls right before making them — aidangomez · 2026-10-27
- Nat Lambert: Cohere's model is narrow RL training on internal infra ill-suited for open weights — natolambert · 2026-10-27
- OpenAI teases DOTS is coming to Europe 'soon' — mark_k · 2026-10-27
- Clifford Pickover challenges AI to decipher a message he encoded in Norse runes — pickover · 2026-10-27